importing-push-tokens
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process device tokens and user IDs from various third-party platforms, which introduces a potential surface for indirect prompt injection attacks.
- Ingestion points: External data enters the agent's workflow via CSV exports or API responses from platforms like OneSignal, Braze, and Customer.io, as outlined in
references/exporting-devices.md. - Boundary markers: There are no instructions for using boundary markers or delimiters to isolate untrusted data fields from the agent's execution context.
- Capability inventory: The skill leverages PostHog-specific tools (
integrations-listandintegrations-push-tokens-create) to perform network operations and data storage. - Sanitization: The instructions lack explicit requirements for sanitizing or validating the imported tokens and IDs to ensure they do not contain hidden instructions or malicious payloads.
Audit Metadata